Surgery is considered when epilepsy is drug-resistant, meaning seizures continue despite adequate trials of two appropriate, tolerated medicines, and when the investigations point consistently to where the seizures begin. A multidisciplinary team then weighs whether that region can be removed or treated safely. When the picture is unclear or the pieces disagree, recording from inside the brain with SEEG may be needed first.
Imagine a 25-year-old woman who has had seizures since she was a teenager. She cannot drive, and she has lost jobs. Then she has surgery, and the seizures stop. A year later she is driving, working and living her life.
That is the ending everyone sees. It is not the whole story. Behind it there is a great deal of work, sometimes years of it. This lesson walks through that journey step by step, from the diagnosis of drug-resistant epilepsy to the decision about treatment.
The story starts with one diagnosis: drug-resistant (refractory) epilepsy. Around two in three people with epilepsy become seizure-free with medication. They never need to meet a neurosurgeon, and that is a good thing. The remaining third continue to have seizures.
The International League Against Epilepsy (ILAE) defines drug-resistant epilepsy as the failure of adequate trials of two tolerated, appropriately chosen and used anti-seizure medication schedules to achieve sustained seizure freedom.
Why two? Because once two appropriate medicines have failed, the chance that the next one will bring lasting seizure freedom becomes small, and it keeps falling with each further drug. This is the moment the patient should become visible to the surgical team: not to take over, but to help answer a different question. Not “which drug next?” but “where does this all begin?”
A routine brain MRI is often reported as normal in epilepsy, because the lesion we are looking for can be tiny. So we use a dedicated epilepsy protocol. The one recommended by the ILAE, known as HARNESS-MRI, has three core sequences:
Because the volumes are 3D, they can be reviewed in any plane, and further sequences can be added when needed.
This is much longer than a routine scalp EEG. The patient stays in hospital, often for around a week, with scalp electrodes and a camera recording continuously, so that seizures can be captured and the clinical behaviour (semiology) matched to the EEG. Medication is often reduced to allow seizures to occur in a safe, monitored setting.
Each test is one piece of the puzzle.
All the pieces come to one table: neurologists, neurosurgeons, neurophysiologists, neuroradiologists and neuropsychologists. The meeting has one big question: do all the pieces point to the same place?
When they do, we call the findings concordant. A typical concordant story: the semiology suggests a temporal lobe onset, the EEG shows right temporal spikes, the MRI shows right hippocampal sclerosis, and PET shows right temporal hypometabolism. Everyone agrees on one hypothesis, and the patient may proceed to surgery with a good chance of becoming seizure-free.
Often the reality is less tidy. The EEG may lateralise to one side or fail to localise; the MRI may show a lesion on the other side, or nothing at all; other tests may point elsewhere. We call this discordant. Then we need more precise information from inside the brain.
We may have a hypothesis about where the seizures originate, the epileptogenic zone: the region of cortex responsible for generating seizures, and the minimum amount of tissue that must be removed or destroyed for the patient to become seizure-free. Proving it is another matter.
Stereo-electroencephalography (SEEG) is considered when non-invasive investigations fail to define the seizure onset clearly, particularly when:
Thin depth electrodes, each with multiple recording contacts, are placed through small holes in the skull into the regions under suspicion, usually with robotic or frame-based stereotactic guidance so that each trajectory matches the plan. The patient is then monitored on the ward, often for one to two weeks, until typical seizures are recorded; electrical stimulation through the contacts is frequently used as well.
One key point: we cannot place electrodes everywhere. SEEG tests a hypothesis, usually a primary one and perhaps an alternative. Without a clear hypothesis before implantation, SEEG is unlikely to provide an answer.
The SEEG findings return to the MDT. Broadly, the options are:
So the next time you see a patient who is seizure-free, or much improved, after surgery, remember the story behind it: a puzzle of many pieces, and a team from several specialties working to solve it.
Educational content for clinicians and trainees, reflecting the author's understanding at the date of review. It is not medical advice: if you are a patient, please discuss your care with your own clinical team.
